People vary in how much B12 they "store" in the liver. It can be up to 5 yrs worth, or much less.
B12 is lost daily, thru the bile. Some may be reabsorbed in the intestines but that varies among people too. People with inflammation will not be working as well as normal intestines.
B12 is also lost thru the kidneys, but less so. People with kidney failure or damage sometimes show elevated B12 serum levels when they are not using supplements.
The term fat soluble (hence stored) and water soluble (excreted) is an old definition, that with time has shown to be variable.
For example some pyridoxal from B6 is stored in muscle tissue, for use in sudden muscular activity. But this is not a huge amount or done in quantity. The newer form of Thiamine called benfotiamine is fat soluble and therefore is somewhat stored compared to regular thiamine which is excreted very quickly in the urine.
People vary considerably in their need for vitamins, and how they metabolize them. There is much leeway therefore in statements based on research available on them.
It has been known for years that niacin can increase blood sugar.
Many doctors do know this in diabetic patients.
From rxlist.com: NiaspanClinical Laboratory Abnormalities
...Chemistry: Elevations in serum transaminases [see WARNINGS AND PRECAUTIONS], LDH, fasting glucose, uric acid, total bilirubin, amylase and creatine kinase, and reduction in phosphorus.
Hematology: Slight reductions in platelet counts and prolongation in prothrombin time [see WARNINGS AND PRECAUTIONS].....
Niacin may also give false-positive reactions with cupric sulfate solution (Benedict's reagent) in urine glucose tests.
Here is the Linus Pauling Institute's data:
They typically go into minute detail with studies, and many people find it difficult to understand it all:
. Large doses of nicotinic acid have been observed to impair glucose tolerance, likely due to decreased insulin sensitivity. Impaired glucose-tolerance in susceptible (pre-diabetic) individuals could result in elevated blood glucose levels and clinical diabetes. Elevated blood levels of uric acid, occasionally resulting in attacks of gout in susceptible individuals, have also been observed with high-dose nicotinic acid therapy (34). Nicotinic acid at doses of 1.5 to 5 grams/day has resulted in a few case reports of blurred vision and other eye problems, which have generally been reversible upon discontinuation.
This link at Linus Pauling Institute goes into detail about studies with niacinamide done on type I children, that were promising for improving pancreatic functions. However, the results were not as great as expected.
The link you quoted, Dr. Smith, is a science writer's simplification of those studies.
As I stated, Niacin (nicotinic acid) has been documented to raise blood sugar in some people. This is not new information, its been around for years.
Readers are dependent on the talent of medical writers in translating data to informational documents. Some are better than others.
And when one little factoid appears in studies, it might be misinterpreted and passed down repeatedly. This happened with a compound that is toxic abbreviated GLA, which happens to cause seizures, and it was gathered in searching MedLine, for articles on evening primrose oil with contains GLA (another compound entirely, this one is gamma Linoleic acid), which is benign. For a long time and probably still (I haven't looked recently) evening primrose oil still contains a caution for causing seizures, on laymen sites concerning evening primrose oil.
I found that by accident, doing searches myself over a decade ago.
The same holds true of pharmacy databases, which vary from store to store. Some are pretty up to date, and others lag behind in flagging drug interactions. Not everything on your handout in a pharmacy may be 100% accurate either. They try hard, but sometimes it take YEARS for some updating, there too. It took forever for some SSRI/serotonin interactions to appear on those data bases. I was finding information on Medline long before they appeared for the public.